11

I have worked on few research labs and performed research as a master's student with PhD students and faced similar issues.

It has been papers that I have contributed most of the work for that paper, including producing results, and writing. Yet the PhD student wants to put my name second due to seniority.

When I first got involved I understood that the order of the paper authors is important in showing the amount of contribution. It feels extremely demotivating to know someone else is getting credit for the very hard work I done and makes me want to stop being involved in research.

Questions:

  • What is the best practice to decide the author order for CS/AI/ML related conferences?
  • When everyone has contributed equally and ordering by name, is it the first name or last name you order by?
  • When the order of the authors doesn't match the level of contribution, what is a good way to bring it up and to who?
  • 3
    What else has the PhD student done? If she designed the experiment (do you have "experiments" in CS? I have no idea) and mentored you throughout the process of "producing results" I could see why she would expect to be first author. It does sound like a mismatch, and I'm not saying she's right but your description of who did what is kinda vague – Azor Ahai Oct 29 '18 at 18:03
  • Most of the times, they provide initial research direction, which many times changes after reformulating the hypothesis and going based off results. Other times they provide the data that will be used, and making sense of them is the purpose of the paper. Just to list few examples, it's not a specific event but rather multiple ones I am trying to describe. – user99355 Oct 29 '18 at 18:24
  • Do you write the bulk of the paper? Sometimes the actual code/statistical analysis of the data is only one section of a paper. But if you write, say, intro, related work, experiment and results section, then the paper is mostly your work. – Pam Oct 29 '18 at 21:50
  • 2
    @Pam Yes for my case I have written the bulk of the paper. – user99355 Oct 29 '18 at 22:07
12

Rule #1: Always try to pin these types of things down before performing the work.

Rule #2: Understand that academia (graduate school) is run on an implicit pecking order:

Departmental pressures and requirements > Advisor > PhD Student > Master's Student

This makes complying to Rule #1 very difficult in practice. Every early career researcher runs into the issue you have described in some form. (Not that this makes such practices right). People higher than you in the pecking order will use administrative leverage to get what they want.

Advisor: "I am the first author on this paper." (And if you disagree, you'll have to find a new advisor and lose a year's worth of work).

Department: "Give Professor Smith what she wants or we will not sign off on your degree."

PhD Student: "Collect this data for me. (And then try to convince someone that you actually did the work and not me)."

The order of attribution should be based on the order of contribution in most CS-based fields. Equal contributions are noted by alphabetical order of the authors' surnames, sometimes with a footnote indicating equal contribution.

I would bring the issue up first with the PhD student directly. I would then speak with his/her advisor. The advisor is essentially the court of last resort. An advisor also will be able to help with the navigation of what actually constitutes authorship for each party. At the very least, your own advisor should (hopefully) be able to help you obtain credit for your work in the form of a chapter in your thesis or something.

Going forward, I would be firm in establishing what the authorship expectations will be on any work you do. You will at times be forced to balance fairness in authorship and expediency in obtaining your degree.

I will note that I take the pecking order things into account when I interview applicants for a job. I care much less about author order and much more about what the person actually did for the paper. Thus, if the OP applied to a job with me, I would likely give his research the same weight whether he was first or second author. I fully understand that superiors sometimes take advantage of their underlings.

  • 4
    Good advice for rule #1, but when I have brought up such a thing in the past in the initial conversation of cooperation the response is "Let's first have a paper and then decide", "Producing a paper is such a difficult process, we don't know if we will accomplish" and so on. – user99355 Oct 29 '18 at 18:28
  • 1
    @user99355. Hence Rule #2. Rule #1 is MUCH harder said than done. If someone higher in the pecking order wants to take advantage of you, they likely will. It is then that you have to balance "fairness" and finishing your degree. – Vladhagen Oct 29 '18 at 18:32
  • 5
    @Vladhagen Don't you mean "much easier said than done"? – Azor Ahai Oct 30 '18 at 0:34
  • @AzorAhai Yes. Good point. – Vladhagen Oct 30 '18 at 2:41
11

My answer to all three of your questions: Don't bother too much. Often such short-term unfair situations will level out.

Next year you start your PhD and write a second paper with the same co-author. And then time you will be the first author and profit a lot from the experience of your co-author. Or someone leaves academia, you have to take over and just by adding the finishing touches, you get a great paper. Or you supervise a brilliant masters student winning you another easy paper. Or an office mate thanks you for your thoughtful coffee break discussion by adding you to his latest paper.

Sometimes (this time) you lose, sometimes (maybe next time) you win.

You have to find out and be aware of the authors wanting to trick their co-authors.

  • 1
    I like this perspective. – user99355 Oct 29 '18 at 23:35
  • 1
    Upvote: I agree with "don't bother too much", but for different reasons (academia.stackexchange.com/a/119230/22768) – user2768 Oct 30 '18 at 7:16
  • True, but the coffee break situation is just wishful thinking! In my experience I have been ruthlessly kicked out of papers (after leaving a group) a few times (3-4), placed 2nd despite having done the work a few times (2-3), and invited to join a paper once (which I rejected because it was 2 hours before submission, so yeah, it would have been unfair). Things level out as you get older as you exploit others in an unfair way as they did to you, so I am not really sure it is a good thing. – Anon Oct 30 '18 at 8:26
  • 1
    @Anon maybe you have a tendency to bring the wrong kind of bagels to the coffee break? – jwenting Oct 30 '18 at 12:44
4

Best practice depends on the particular (sub-)field. Look at how it is in yours. Also, rules are flexible.

In mine, the AMS guidelines are usually the norm: alphabetical family name ascending lexicographic order, regardless of authors' individual contributions.

In any case, you may explicitly state contribution in a footnote.

Other comments stress clearly enough the importance of making things clear ahead of writing. I would just add that your own contributions will speak for themselves regardless of the order in which you appear as an author (as long as you do!).

3

I agree with the don't bother too much philosophy [in response to all questions], but I disagree with the motivation ("unfair situations will level out"), I believe motivation should come from it not really mattering: You're career is built on your research, not author ordering.

Do great work; don't worry about author ordering.


This advice may not apply to all disciplines.

  • "Your career" might get stuck at some early point, if ALL you can show is 2nd author publications. – BPND Oct 30 '18 at 8:15
  • 1
    @BPND Do you have non discipline-specific examples? Or even discipline specific examples? – user2768 Oct 30 '18 at 9:34
  • Out of curiosity, how can any example be non discipline-specific? – Mark Meckes Oct 30 '18 at 16:20
  • @MarkMeckes If it applies across multiple disciplines? – user2768 Oct 30 '18 at 16:41
  • @user2768 This was partially covered in another question on this site. Further, in an analysis of publication-career correlations, the authors find that later PIs have a two-fold increase in their first-author publication rate compared to authors who do not become PIs. Obviously, all this does not matter for fields that exclusively order alphabetically. – BPND Nov 2 '18 at 12:41
2

I disagree with the other answers: to me, this is some clearly unethical behaviour from your co-author and it does matter, especially if this makes you want to stop doing research. While this can happen of course, in my experience this is quite rare (I'm also in the field of AI/ML). I would encourage you to raise the issue with your co-authors.

What is the best practice to decide the author order for CS/AI/ML related conferences?

In CS/AI/ML the authors are almost always ordered by the importance of their respective contribution, and the reader is likely to interpret it this way.

When everyone has contributed equally and ordering by name, is it the first name or last name you order by?

Last name, but I saw this only once in a multidisciplinary paper.

When the order of the authors doesn't match the level of contribution, what is a good way to bring it up and to who?

First with the co-author directly, it might be a simple misunderstanding. Maybe this person can give you a good reason why they should be first author: did they have to rewrite everything after you? Did they fix major mistakes? Did they contribute major elements of the work? If not, this has to be discussed with the other persons involved, including the advisor. If you are not comfortable talking to the advisor directly, you might talk about it to your director of studies or any professor who could tell you if they think that this is acceptable or not and, if not, address the issue themselves with the advisor.

  • 3
    In more theoretical subfields of CS/ML, authors are always ordered alphabetically. – JeffE Oct 30 '18 at 2:15
  • @JeffE I searched Google Scholar for machine learning and the second hit didn't use alphabetical ordering, nor did the third or fifth (the fourth was single author). Even in subfields, I doubt that every manuscript complies. – user2768 Oct 30 '18 at 8:12
  • 1
    @user2768 That’s why I wrote “In more theoretical corners of...”. ML is a huge field, with many subcommunities. – JeffE Oct 30 '18 at 11:26
  • Order by the surname or family name, not the given name. In the US, then, order by the last name. But elsewhere names may not be ordered like this. Mao Zedung is ordered under "M". – GEdgar Oct 30 '18 at 13:17
  • 1
    One should keep in mind, though, that "ordered by the importance of their respective contribution" is not necessarily the same as "ordered by the time they spent on it". Coming up with the right idea (3 days) may or may not be more important than implementing it (3 months). – Uwe Oct 30 '18 at 13:44
0

When climbing a ladder, sometimes helping another is a way forward to your own achievement further down the line. I would posit, that while you made a sizeable contribution - was this your first paper or 12th? If it was your first paper I somehow doubt that your contribution should outweigh anyone else's who have more experience than you? Nonetheless, if you have been slighted, the very inclusion as an author should be a big boost. In any event, a footnote or endnote indicating your actual contribution should rectify any grievance you might have, and satisfy the pecking order, which often has little to do with reality.

0

What is the best practice to decide the author order for CS/AI/ML related conferences?

The best practice is to list authors randomly (or alphabetically), and when necessary, explicitly state that author name order does not represent degree of contribution.

This is not to say it is the prevalent practice in all fields. The custom of somehow assigning more significance based on name order should be rooted out, and it will not be rooted out until enough people endeavor to root it out.

While you might not be able to make this happen right now - having perhaps the least amount of "political clout", you can still offer this principle to the PhD candidate as a compromise, instead of you two quarreling about who is more deserving of being first.

When everyone has contributed equally and ordering by name, is it the first name or last name you order by?

In European-type names (which the last name is the family name) - last name first.

When the order of the authors doesn't match the level of contribution, what is a good way to bring it up and to who?

The order of authors very frequently doesn't match the level of contribution, even when it supposedly should. But I don't want to evade the question. First, bring it up among all authors (e.g. via email or group meeting). Try to make sure you have enough backing in written evidence to establish your having the most significant contribution, otherwise "that's just, like, your opinion man". If you have enough backing among the authors but get overruled by your advisor, you may want to consider taking this up with your graduate researchers' union.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.